Data Cleaning Prompt

Clean and normalize this data:

RAW DATA:
[The messy data]

CLEANING OPERATIONS:
1. Standardize: [what to standardize, e.g., "phone numbers to E.164"]
2. Normalize: [what to normalize, e.g., "company names (remove Inc., LLC)"]
3. Fix: [common errors to correct, e.g., "obvious typos in country names"]
4. Remove: [what to strip, e.g., "extra whitespace, HTML tags"]
5. Parse: [what to extract, e.g., "first name and last name from full name"]

OUTPUT FORMAT:
{
  "cleaned_data": { ... },
  "changes_made": [
    {
      "field": "field_name",
      "original": "value",
      "cleaned": "value",
      "operation": "what was done"
    }
  ],
  "issues_found": [
    {
      "field": "field_name",
      "issue": "description",
      "action_taken": "what was done"
    }
  ]
}

PRESERVATION:
- Keep original field values in separate object if significantly changed
- Don't discard data unless explicitly instructed